Saturday, February 25, 2017

Not Going as Smoothly As Hoped

I've been working on the enclosure for my mill since the post last Saturday, and it isn't exactly going effortlessly.  Smooth isn't a word that I'd use.  I spent a lot of time verifying dimensions and then cut the aluminum extrusion to length, then cut all the opaque plastic panels.  Finally, I cut the two wider transparent panels so that I could build the side panels.  These seemed like an easy place to start. 

I have the two side panels built.  These are one opaque and one transparent panel in a frame with a divider between them.  The rendering looks like this
The way I tried to build these was to build the top and bottom rails first.  These are 35" pieces of the 15mm extrusion this is built from.  They get stamped steel L brackets which require four M3 5mm screws and M3 nuts, and I mounted the L brackets that go in the ends, tightening only the two screws that anchor the bracket to the top or bottom rails.  I left the bracket in the middle loose until I got the plastic panels in the right places.  I put a few (~2 to 4) extra nuts in each segment in case I want to add something at a later date.  The L brackets attach the three uprights to the top and bottom rails.  It might be better at this point to show you a finished panel so that maybe you can see these details. 
The main difficulty is that the white corrugated plastic (Coroplast) is 4mm thick nominal, while the groove in the extrusion is 3mm wide.  The "all you gotta do" answer is to slit the ends of the corrugations and the length of the sides, and then it slides together easily.  Not quite.  My idea was to slide the white plastic panel into the slot in the extrusion along the right vertical piece, into the groove in the bottom and then slide the middle upright onto the panel.  That was troublesome.  Because the plastic is being compressed into place, it works better if you slide the extrusion along the edge of the Coroplast, except that it gets a bit harder to slide it every inch as the friction increases.  Getting the entire edge in place at once turns out to be essentially impossible, but with patience, pressure, and small jeweler's screwdriver to help center the panel, it can be forced into the groove all the way around.  Once the white half was made, I slid the acrylic into the left side, and that was enormously easier.  It's 2mm thick in a 3mm slot.  Even with the corrugations slit to make the Coroplast more compressible, it's rough to get in the slot. 

The second one took a bit less time, but was still rough to do.  I did most of it with the pieces lying horizontally on a convenient flat surface (our freezer) and alternated between that and standing it on edge on the floor.  Mrs. Graybeard was a great help and did  a lot of the fussy details of assembly.  In places where my height and strength advantage would help, I took over.  

The back panel is going to be rough without building some sort of fixture to help assembly, or some other changes.  The most direct would be to widen that 3mm slot so that the 4mm plastic fits into it more easily.  I haven't worked out just how to do it on any machine, but I have a 3/16" end mill, which would make the slot 4.76 mm wide.  Widening the slot would make assembly much easier. 


  1. Could you lightly sand the edge of the Coroplast?

    I'd probably just take my Dremel to the channel in the extrusion and open it up a bit.

    Perhaps a little silicone spray lube would help....

  2. > The most direct would be to widen that 3mm slot so that the 4mm plastic fits into it more easily.

    If you're crazy, clamp down large wooden guide blocks to deal with the greater cutting forces, and wear ear protection, you can saw aluminum on a wood table saw with a carbide blade. You don't have to take all the depth off in one pass.

    1. That's currently the most likely option. Second most likely is my router table, but I don't think I have a bit that's undersized so I could take off a few mils at a time.

      Both scare me but I think the router is scarier.

      There's some saying like "sure you're scared, but are you scared enough?"

  3. How much wall thickness does the extrusion have? Or is it solid?
    15mm is a skosh less than 5/8", yes? That should run OK on a table saw, watch the blade/insert clearance and use a low hook blade. It will make a mess of aluminum chips, a hell of a racket, and don't get in front of the rip- it will make an extremely effective arrow if you get a kickback. Use a good push stick, one that will hold down the stock onto the table,so it does not bounce around, and with a broad enough hook on the end so there is no chance of the push stick getting between the fence and the stock.

    Can't you determine location and press the plastic into position instead of sliding it? Does it have any memory at all? Using some sort of improvised tool to pre-compress the edge of the plastic by bending it in might assist in feeding it smoothly into the extrusion. Same idea as taking a slight bevel cut on a tight fitting plywood panel edge, to help it get in the dado.

    1. The extrusion is shown here The slot is 3mm wide at the top but opens up after just .043". You can see I have very little material to cut off, just the edges of the thin part that make it too tight. The wide part underneath that is 0.225" wide, so the 4mm panel is loose in it.

      I've done some experiments that make it look like I can cut the short (23-5/8) pieces on my big mill. I have a 3/16 end mill and it cut like a hot knife through butter. I cut as much as I can, flip it end for end and finish the cut. I don't know if I can cut the long pieces with it, those are 64" long and I need to experiment with what I can get on the table. I still might need to use the table saw. It's a light cut, like .025" on a side and I don't think it will be difficult at all. IIRC, the teeth on the saw blade are .062. A third of a tooth on each side?


  4. You could run it on the saw. If the slot has to be centered, cut both sides. Otherwise just one. Or run it on the mill and keep sliding the part- a cool trick for this is to use dowel pins that are a press fit in the table tee grooves-3/8", 1/2", 5/8", whatever the groove size is. Tap them in and use them as an index as the part is moved and re-clamped.
    4mm is .1575" The called out spec on the groove is 3.4mm, or .1339" So the difference is .1575-.1339= .0236" or divide by two to get the cut on each side- .0118" Obviously all these numbers are better suited to fine tolerance rather than cramming in a plastic sheet! And the spec may be different than the metal, you said it was 3mm. I am more inclined to believe your measurement than the spec!

    A hint- I suggest you work either in metric or inches- conversion can really be a zone fraught with danger. A wise guy mentioned this to me when I asked for comments on a drawing I made, with both metric and inch measurements on it. Isn't that how the Hubble got screwed up the first time?

    1. I do work in inches, unless I have to do otherwise. It wasn't Hubble, it was the Mars Climate Orbiter. They got to the final course correction and never heard from the probe again, because of a metric conversion that wasn't done. (Hubble was a different error - as an amateur optician, I can fill a column on that)

      I just cut the first 64" rail. I squared my mill's vise as well as I could get it, and then cut the 64" 16" at a time. I just cut 16", moved the mill back to the starting point, slid the rail in the vise to get another 16" ready to cut and repeated. After the third cut, I turned the rail end-for-end and cut the last 16". It came out fine. The Coroplast is much easier to work with in this rail. I should be able to do all the rest tomorrow.

  5. So something else got the Hubble. Thank you for the clarification.

    Just curious here, I am trying to picture cutting 16" on one pass with the stock held in a vise. \Wasn't there an awful lot of material hanging out past the vise jaws?

    1. It was definitely a "circus trick" cut, and I'd hesitate to call it machining, but it didn't need high accuracy. What I got was on the order of DrJim up there saying to shave it back with a Dremel, but much faster.

      The vise was at zero, but not exactly centered, and I cut from -8 to +8 on the table. The vise clamping surfaces are only 3" wide. At one point, there was about 56 inches on a side that stuck out well beyond the enclosure. That was supported with blocks to keep the extrusion from trying to lift out of the vise, but it wasn't clamped.

      Sometimes you gotta do what you gotta do.